A relatively high percentage of children are born with prenatal exposure to cocaine and research suggests that this may retard fetal growth and have a negative effect on the neurobehavioral development of the newborn. This suggests that prenatal cocaine exposure may place children at risk for later cognitive development. However, there is a paucity of data concerning the later development of these children, particularly in the area of language. In the only preliminary and limited study on language development, impairment has been reported. An investigation of the language development of these children has important theoretical implications in terms of the effects prenatal cocaine exposure has on brain development and later language function. It also has important clinical implications in terms of the early identification and treatment of language impairment in this population. The primary objective of the proposed research is to assess the language development of young children prenatally exposed to cocaine. The goal of this preliminary research is to provide pilot data on this population and to evaluate the sensitivity of the methods for use in a prospective study. After 2 years, a subsample of the children will be followed to validate the preliminary findings. This data will facilitate the early identification of at risk children so that treatment effectiveness can be maximized. The language development of a group of 25 children with prenatal exposure to cocaine will be compared to a matched cocaine-free control group of 25 subjects. All subjects will be within the age range of 18-24 months. The language development of the two groups will be compared through the use of established measures of language production and comprehension that provide an assessment of the pragmatic, semantic, and form components of language. The language production measures will be based on the analysis of a spontaneous language sample. The language comprehension measures will be based on the administration of two structured comprehension tasks. The overall approach to data analysis will be to perform a series of multivariate discriminant analyses. Based on the results of these analyses, a final discriminant model will be developed.